ThreadPoolExecutor提供了四个构造方法：

ThreadPoolExecutor构造方法.png

我们以最后一个构造方法（参数最多的那个），对其参数进行解释：

 public ThreadPoolExecutor(int corePoolSize, // 1
                              int maximumPoolSize,  // 2
                              long keepAliveTime,  // 3
                              TimeUnit unit,  // 4
                              BlockingQueue<Runnable> workQueue, // 5
                              ThreadFactory threadFactory,  // 6
                              RejectedExecutionHandler handler ) { //7
        if (corePoolSize < 0 ||
            maximumPoolSize <= 0 ||
            maximumPoolSize < corePoolSize ||
            keepAliveTime < 0)
            throw new IllegalArgumentException();
        if (workQueue == null || threadFactory == null || handler == null)
            throw new NullPointerException();
        this.corePoolSize = corePoolSize;
        this.maximumPoolSize = maximumPoolSize;
        this.workQueue = workQueue;
        this.keepAliveTime = unit.toNanos(keepAliveTime);
        this.threadFactory = threadFactory;
        this.handler = handler;
    }

如果对这些参数作用有疑惑的请看 ThreadPoolExecutor概述。
知道了各个参数的作用后，我们开始构造符合我们期待的线程池。首先看JDK给我们预定义的几种线程池：

一、预定义线程池

1. FixedThreadPool

    public static ExecutorService newFixedThreadPool(int nThreads) {
        return new ThreadPoolExecutor(nThreads, nThreads,
                                      0L, TimeUnit.MILLISECONDS,
                                      new LinkedBlockingQueue<Runnable>());
    }

• corePoolSize与maximumPoolSize相等，即其线程全为核心线程，是一个固定大小的线程池，是其优势；
• keepAliveTime = 0 该参数默认对核心线程无效，而FixedThreadPool全部为核心线程；
• workQueue 为LinkedBlockingQueue（无界阻塞队列），队列最大值为Integer.MAX_VALUE。如果任务提交速度持续大余任务处理速度，会造成队列大量阻塞。因为队列很大，很有可能在拒绝策略前，内存溢出。是其劣势；
• FixedThreadPool的任务执行是无序的；

适用场景：可用于Web服务瞬时削峰，但需注意长时间持续高峰情况造成的队列阻塞。

2. CachedThreadPool

     public static ExecutorService newCachedThreadPool() {
        return new ThreadPoolExecutor(0, Integer.MAX_VALUE,
                                      60L, TimeUnit.SECONDS,
                                      new SynchronousQueue<Runnable>());
    }

• corePoolSize = 0，maximumPoolSize = Integer.MAX_VALUE，即线程数量几乎无限制；
• keepAliveTime = 60s，线程空闲60s后自动结束。
• workQueue 为 SynchronousQueue 同步队列，这个队列类似于一个接力棒，入队出队必须同时传递，因为CachedThreadPool线程创建无限制，不会有队列等待，所以使用SynchronousQueue；

适用场景：快速处理大量耗时较短的任务，如Netty的NIO接受请求时，可使用CachedThreadPool。

3. SingleThreadExecutor

    public static ExecutorService newSingleThreadExecutor() {
        return new FinalizableDelegatedExecutorService
            (new ThreadPoolExecutor(1, 1,
                                    0L, TimeUnit.MILLISECONDS,
                                    new LinkedBlockingQueue<Runnable>()));
    }

咋一瞅，不就是newFixedThreadPool(1)吗？定眼一看，这里多了一层FinalizableDelegatedExecutorService包装，这一层有什么用呢，写个dome来解释一下：

    public static void main(String[] args) {
        ExecutorService fixedExecutorService = Executors.newFixedThreadPool(1);
        ThreadPoolExecutor threadPoolExecutor = (ThreadPoolExecutor) fixedExecutorService;
        System.out.println(threadPoolExecutor.getMaximumPoolSize());
        threadPoolExecutor.setCorePoolSize(8);
        ExecutorService singleExecutorService = Executors.newSingleThreadExecutor();
//      运行时异常 java.lang.ClassCastException
//      ThreadPoolExecutor threadPoolExecutor2 = (ThreadPoolExecutor) singleExecutorService;
    }

对比可以看出，FixedThreadPool可以向下转型为ThreadPoolExecutor，并对其线程池进行配置，而SingleThreadExecutor被包装后，无法成功向下转型。因此，SingleThreadExecutor被定以后，无法修改，做到了真正的Single。
使用场景：
希望多个任务排队执行。线程数固定为 1，任务数多于 1 时，会放入无界队列排队。任务执行完毕，这唯一的线程也不会被释放

4. ScheduledThreadPool

    public static ScheduledExecutorService newScheduledThreadPool(int corePoolSize) {
        return new ScheduledThreadPoolExecutor(corePoolSize);
    }

newScheduledThreadPool调用的是ScheduledThreadPoolExecutor的构造方法，而ScheduledThreadPoolExecutor继承了ThreadPoolExecutor，构造是还是调用了其父类的构造方法。

    public ScheduledThreadPoolExecutor(int corePoolSize) {
        super(corePoolSize, Integer.MAX_VALUE, 0, NANOSECONDS,
              new DelayedWorkQueue());
    }

整个线程池表现为：线程数固定，任务数多于线程数时，会放入无界队列排队。任务执行完毕，这些线程也不会被释放。用来执行延迟或反复执行的任务

二、自定义线程池

以下是自定义线程池，使用了有界队列，自定义ThreadFactory和拒绝策略的demo：

public class ThreadTest {
    public static void main(String[] args) throws InterruptedException, IOException {
        int corePoolSize = 2;
        int maximumPoolSize = 4;
        long keepAliveTime = 10;
        TimeUnit unit = TimeUnit.SECONDS;
        BlockingQueue<Runnable> workQueue = new ArrayBlockingQueue<>(2);
        ThreadFactory threadFactory = new NameTreadFactory();
        RejectedExecutionHandler handler = new MyIgnorePolicy();
        ThreadPoolExecutor executor = new ThreadPoolExecutor(corePoolSize, maximumPoolSize, keepAliveTime, unit,
                workQueue, threadFactory, handler);
        executor.prestartAllCoreThreads(); // 预启动所有核心线程
        for (int i = 1; i <= 10; i++) {
            MyTask task = new MyTask(String.valueOf(i));
            executor.execute(task);
        }
        System.in.read(); //阻塞主线程
    }
    static class NameTreadFactory implements ThreadFactory {
        private final AtomicInteger mThreadNum = new AtomicInteger(1);
        @Override
        public Thread newThread(Runnable r) {
            Thread t = new Thread(r, "my-thread-" + mThreadNum.getAndIncrement());
            System.out.println(t.getName() + " has been created");
            return t;
        }
    }
    public static class MyIgnorePolicy implements RejectedExecutionHandler {
        public void rejectedExecution(Runnable r, ThreadPoolExecutor e) {
            doLog(r, e);
        }
        private void doLog(Runnable r, ThreadPoolExecutor e) {
            // 可做日志记录等
            System.err.println( r.toString() + " rejected");
//          System.out.println("completedTaskCount: " + e.getCompletedTaskCount());
        }
    }
    static class MyTask implements Runnable {
        private String name;
        public MyTask(String name) {
            this.name = name;
        }
        @Override
        public void run() {
            try {
                System.out.println(this.toString() + " is running!");
                Thread.sleep(3000); //让任务执行慢点
            } catch (InterruptedException e) {
                e.printStackTrace();
            }
        }
        public String getName() {
            return name;
        }
        @Override
        public String toString() {
            return "MyTask [name=" + name + "]";
        }
    }
}

输出结果如下：

image.png
其中线程线程1-4先占满了核心线程和最大线程数量，然后4、5线程进入等待队列，7-10线程被直接忽略拒绝执行，等1-4线程中有线程执行完后通知4、5线程继续执行。
总结，通过自定义线程池，我们可以更好的让线程池为我们所用，更加适应我的实际场景。

作者：徐志毅
链接：https://www.jianshu.com/p/f030aa5d7a28
来源：简书